Policy Decision Making on Synthetic d_phi=2 out-sample
5.64Policy Error Rate (ER_out)InvTARNet
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| InvTARNetRepresentation dimension (d_phi)=2, Number of training samples (ntrain)=1,000, Delta parameter (δ)=0.00052023.11 | 5.64 | -0.02 | |
| CFR-ISW (WM; α = 1.0)Representation dimension (d_phi)=2, Number of training samples (ntrain)=1,000, Delta parameter (δ)=0.00052023.11 | 7.27 | -1.86 | |
| BWCFR (WM; α = 1.0)Representation dimension (d_phi)=2, Number of training samples (ntrain)=1,000, Delta parameter (δ)=0.00052023.11 | 7.44 | -4.57 | |
| RCFR (WM; α = 1.0)Representation dimension (d_phi)=2, Number of training samples (ntrain)=1,000, Delta parameter (δ)=0.00052023.11 | 8 | -4.27 | |
| k-NNNumber of training samples (ntrain)=1,000, Delta parameter (δ)=0.00052023.11 | 8.18 | — | |
| TARNetRepresentation dimension (d_phi)=2, Number of training samples (ntrain)=1,000, Delta parameter (δ)=0.00052023.11 | 9.82 | -3.73 | |
| CFR (WM; α = 1.0)Representation dimension (d_phi)=2, Number of training samples (ntrain)=1,000, Delta parameter (δ)=0.00052023.11 | 10.88 | -7.97 | |
| CFR (MMD; α = 0.1)Representation dimension (d_phi)=2, Number of training samples (ntrain)=1,000, Delta parameter (δ)=0.00052023.11 | 11.92 | -5.54 | |
| CFR (WM; α = 2.0)Representation dimension (d_phi)=2, Number of training samples (ntrain)=1,000, Delta parameter (δ)=0.00052023.11 | 13.19 | -6.28 | |
| C-ForestNumber of training samples (ntrain)=1,000, Delta parameter (δ)=0.00052023.11 | 16.1 | — | |
| BNN (MMD; α = 0.1)Representation dimension (d_phi)=2, Number of training samples (ntrain)=1,000, Delta parameter (δ)=0.00052023.11 | 16.15 | -4.19 | |
| BARTNumber of training samples (ntrain)=1,000, Delta parameter (δ)=0.00052023.11 | 17.37 | — | |
| CFR (MMD; α = 0.5)Representation dimension (d_phi)=2, Number of training samples (ntrain)=1,000, Delta parameter (δ)=0.00052023.11 | 17.89 | -7.27 |